SeineSailor

Overview

SeineSailor is an AI-powered copilot for project managers, designed to assist with code reviews, project discussions, and knowledge transfer. It leverages the power of llama3 and mixtral-8x7b models, hosted by Cloud providers such as IBM WatsonX, to provide intelligent and context-aware support throughout the development lifecycle.

SeineSailor is inspired by many research works [1, 2] and open source projects[3, 4]

Features

Here's the content formatted as a Markdown table:

Feature	Description
PR Review and Summarization	SeineSailor provides AI-powered automated reviews and summaries for pull requests on demand. It analyzes the changes and offers insights to streamline the review process.
Code change suggestions	Reviews the changes and provides code change suggestions.
Project Discussion Participation	Engage SeineSailor in discussions across various forums including pull request review comments, issues, and discussions by mentioning @SeineSailor. It can contribute to brainstorming, provide clarifications, summarize discussions, suggest code changes, etc.
Knowledge Transfer and Query Resolution	Leverages a comprehensive knowledge base to assist developers with queries about the codebase, architectural design, and specific functionalities, aiding in faster onboarding and information sharing.
Smart review skipping	By default, skips in-depth review for simple changes (e.g., typo fixes) and when changes look good for the most part. It can be disabled by setting review_simple_changes and review_comment_lgtm to true.
Customizable prompts	Tailor the system_message, summarize, and summarize_release_notes prompts to focus on specific aspects of the review process or even change the review objective.

Installation

Add the following file to your repository at .github/workflows/seinesailor.yml:

name: SeineSailor

permissions:
  contents: read
  pull-requests: write
  issues: write
  discussions: write

on:
  issues:
    types: [ opened ]
  issue_comment:
  discussion:
  discussion_comment:
  pull_request_review_comment:
    types: [ created ]
  pull_request_target:
    types: [ opened, synchronize, reopened, edited ]

concurrency:
  group:
    ${{ github.repository }}-${{ github.event.number || github.head_ref ||
    github.sha }}-${{ github.workflow }}-${{ github.event_name }}
  cancel-in-progress: ${{ github.event_name == 'pull_request_target' }}
  
jobs:
  seinesailor:
    if: |
      github.event_name == 'pull_request_target' ||
      (contains(github.event.comment.body, '@SeineSailor') && (
        github.event_name == 'issue_comment' || 
        github.event_name == 'pull_request_review_comment' ||
        github.event_name == 'discussion_comment')) ||
      (github.event_name == 'issues' && contains(github.event.issue.body, '@SeineSailor')) ||
      (github.event_name == 'discussion' && contains(github.event.discussion.body, '@SeineSailor'))
    runs-on: ubuntu-latest
    steps:
      - name: Check for mention outside of quotes in comments
        id: check_mention
        env:
          COMMENT_BODY: ${{ github.event.comment.body }}
        run: |
          echo "checking for mention outside of quotes and code blocks"
          mention_outside_quotes=false
          in_code_block=false
          while IFS= read -r line; do
            # Check for start or end of a code block
            if [[ "$line" =~ ^\`\`\` ]]; then
              if [[ "$in_code_block" = true ]]; then
                in_code_block=false
              else
                in_code_block=true
              fi
            fi
        
            # Process lines that are not in a code block or quoted
            if [[ "$in_code_block" = false && ! "$line" =~ ^\> ]]; then
              if echo "$line" | grep -q '@SeineSailor'; then
                mention_outside_quotes=true
                break
              fi
            fi
          done <<< "$COMMENT_BODY"
          echo "mention_outside_quotes=$mention_outside_quotes" >> $GITHUB_ENV
      - name: Dump event payload for debug
        run: cat $GITHUB_EVENT_PATH
      - name: Run SeineSailor
        uses: SeineAI/SeineSailor@main # or release version tags
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          IBM_CLOUD_API_KEY: ${{ secrets.IBM_CLOUD_API_KEY }}
          WATSONX_PROJECT_ID: ${{ secrets.WATSONX_PROJECT_ID }}
        with:
          debug: false

Environment Variables

• GITHUB_TOKEN: This should already be available in the GitHub Action environment. It is used to interact with the GitHub API.
• IBM_CLOUD_API_KEY and WATSONX_PROJECT_ID: Use these to authenticate with the WatsonX API. Please add this key to your GitHub Action secrets.
• OPENAI_API_KEY: If you want to use OpenAI models instead. Please add this key to your GitHub Action secrets, and set in the last part of yaml:

        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
        with:
          debug: false
          tag: stable
          llm_api_type: openai
          llm_light_model: gpt-3.5-turbo
          llm_heavy_model: gpt-4
          api_base_url: https://api.openai.com/v1

Usage

To summon SeineSailor, simply mention @SeineSailor in a comment within a pull request, issue, or discussion. SeineSailor will analyze the context and provide a response based on its understanding of the codebase and project history.

Examples

• Requesting a PR review:

  @SeineSailor Please review this pull request and provide a summary of the changes.

• Asking for clarification on a function:

  @SeineSailor Can you explain how the processData function works and how it interacts with the rest of the codebase?

• Seeking guidance on architecture:

  @SeineSailor What is the overall architecture of this project? How are the different components interconnected?

Contributing

Contributions to SeineSailor are welcome! If you have any suggestions, bug reports, or feature requests, please open an issue on the GitHub repository.

Important! Please read Contributing_README.md to understand how we separate production and development process.

License

SeineSailor is released under the MIT License.